Draft data. Specs and scores are compiled placeholders, not vendor-verified. How we score.
News · Research

Figure 03 for home use and scaling — evidence check

Source: BMW Group

Figure 03 Figure AI humanoid robot

Lead strip: Status: Pilot (BMW Spartanburg sequencing) + vendor scale/home demos · Availability: enterprise partnership; consumer list price — · All-in cues: CapEx/RaaS — · Primary: /robots/figure-03/ · Scoring: methodology

Evidence brief only — not a sold ranking. Figure 02 history ≠ Figure 03 current.

Figure AI’s third-generation humanoid sits at the intersection of three storylines that readers often blur: BMW factory evidence, BotQ production scaling claims, and home-task demonstrations under Helix 02. This check keeps those lanes separate and points to the product page: Figure 03.

Robot strips (context, not a trophy board)

  • Figure 03 · Figure AI · Status: Pilot (BMW Spartanburg sequencing) · Use: factory logistics · Evidence: BMW Group + Figure primary posts · → /robots/figure-03/
  • Digit · Agility Robotics · Status: Digit 4 field / Digit 5 announced · Use: warehouse logistics · Evidence: named sites + published hours/totes · → /robots/digit/
  • Optimus · Tesla · Status: Research / production-ramp · Use: factory R&D / Academy · Evidence: Tesla IR; external buy page — · → /robots/optimus/

Rule one: Figure 02 ≠ Figure 03

BMW Group’s 25 June 2026 release states that after a successful deployment with Figure 02 at Plant Spartanburg, the further-developed successor Figure 03 will start on complex sequencing applications in logistics. Ulrich Wieland (BMW Manufacturing) called Spartanburg the “birthplace of humanoid robotics” in BMW’s day-to-day operations and noted the completed Figure 02 body-shop pilot. Brett Adcock (Figure) described an 11-month Figure 02 deployment proving humanoids can be assets on the floor — language about the prior generation (BMW Group).

BMW also states Figure 02 supported production of more than 30,000 BMW X3 vehicles over ten months in the body shop (sheet-metal insertion for welding). That record belongs to Figure 02.

Figure’s own 30 June 2026 post places Figure 03 in Hall 52 on a sequencing workflow — picking unsorted parts, placing them into sequencing trolleys, and whole-body cart pulling under Helix 02 — explicitly transitioning “from classic pick-and-place sheet metal loading” (Figure: F.03 at BMW).

Do not assign Figure 02’s hours, vehicle counts, or body-shop tenure to Figure 03. Different generation, different workflow. Criteria: methodology.

What BMW evidence actually covers for Figure 03

Documented (partner + vendor):

  • Figure 03 starts / arrives for logistics sequencing at Spartanburg (BMW + Figure).
  • Use case: unsorted parts → sequencing trolley → collection point for just-in-sequence delivery (BMW process description).
  • Helix 02 loco-manipulation claims for simultaneous manipulation and cart pulling (Figure).

Not documented in those releases:

  • Robot count on the Spartanburg floor for Figure 03 (— — do not invent a fleet size).
  • Cycle-time SLAs, intervention rates, or multi-year RaaS terms for Figure 03 (—).
  • Consumer home availability tied to the BMW project (—; this is a factory logistics project).

Home-use claims: demos ≠ household deployment

Figure’s Helix 02 materials (27 January 2026) describe a continuous ~4-minute dishwasher loco-manipulation demonstration across a full-sized kitchen — walking, unloading/loading, cabinet stacking — framed as autonomous with 61 loco-manipulation actions and no human intervention in the published demo narrative (Figure: Helix 02). The same post highlights Figure 03 hardware features (palm cameras, tactile sensing) used in dexterity demos.

How to read this for buyers:

  • Strong lab / demo evidence of long-horizon home-like tasks on company video.
  • Not the same as named household customers, warranty for residential use, or a public consumer list price (—).
  • BotQ allocation language later mentions robots going to “end-to-end housework” development among other buckets — that is R&D allocation, not a retail home SKU (Figure: Ramping Figure 03 Production).

Home autonomy remains a claim to test against real dwellings, safety regimes, and support models — separate from BMW Hall 52.

Further reading: which home chores robots can realistically handle by 2030 and 2035, a capability map from our parent company House of 47.

Scaling claims: BotQ production vs field fleets

Figure’s 29 April 2026 production post claims BotQ delivered over 350 third-generation robots and moved from 1 robot per day to 1 per hour (~24×) in under 120 days, with end-of-line first-pass yield “over 80%,” battery-line first-pass yield 99.3%, 9,000+ actuators, 50+ in-process inspection points, and 80+ functional tests per robot (Figure).

Figure also states robots ship to internal R&D, data collection, housework development, and commercial use-case development, and that service tooling extends to HQ, customer sites, and residential homes.

Evidence discipline:

Claim typeWhat it supportsWhat it does not prove
BotQ cadence / yield (vendor)Manufacturing scale effortMulti-customer paid fleet hours
350+ units delivered (vendor)Hardware volume for data + pilotsThat all units are in customer production
BMW Figure 03 sequencingPartner-stated factory projectHome product launch
Helix 02 dishwasher demoAutonomy demo qualityConsumer buyability
Figure 02 BMW 30k vehiclesPrior-generation field historyFigure 03 tenure

Compare manufacturing-scale narrative with logistics field depth on Digit, and with IR-only buyability limits on Optimus. We do not convert vendor unit counts into a sold “#1 scaler” ranking.

All-in cost cues

Public consumer list price for Figure 03: —. Enterprise commercials are partnership-specific. Planners should budget integration, safety case, supervision, spares, and facility redesign separately from any informal valuation headlines. Unknown terms stay —.

Practical takeaway

  1. Credit Figure 03 only with Figure 03 evidence — BMW sequencing + Helix 02 demos + BotQ claims as labeled.
  2. Keep Figure 02’s 30,000-vehicle / 10–11 month body-shop record in the history column.
  3. Home demos ≠ home SKU; ask for residential support, safety, and price before treating housework videos as procurement.
  4. Cross-read Figure 03, Digit, Optimus, and methodology.

FAQ

Is Figure 03 available for home purchase?

No public consumer checkout or list price appears in the BMW or Figure materials cited here. Home-task videos and BotQ “housework development” allocations are not a retail SKU. Availability for households: —.

Did Figure 03 assemble 30,000 BMW X3s?

No. BMW attributes the more than 30,000 X3 / ~ten-month body-shop support to Figure 02. Figure 03 is described for logistics sequencing. Figure 02 ≠ Figure 03.

How strong is the scaling evidence?

Figure publishes unusually detailed BotQ cadence and yield claims. Treat them as vendor manufacturing evidence. They do not automatically equal multi-site paid operational hours like Digit’s logistics narrative, and they do not replace BMW’s partner statement for the Spartanburg Figure 03 project.

How should Figure 03 sit next to Digit and Optimus?

Digit: longer published warehouse operational narrative. Optimus: Research / IR production-ramp without a public buy page. Figure 03: BMW partner-stated sequencing pilot plus vendor scale/home demos. Different evidence classes — not a trophy board.

Sources

Figure 03 robot count at BMW, consumer price, and residential delivery terms: —.